Carbon Dioxide in a Supported Ionic Liquid Membrane: Structural and

Subscriber access provided by UNIV OF NEWCASTLE

Article

Carbon Dioxide in a Supported Ionic Liquid Membrane: Structural and Rotational Dynamics Measured with 2D IR and Pump-Probe Experiments Jae Yoon Shin, Steven A Yamada, and Michael D. Fayer J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b05759 • Publication Date (Web): 19 Jul 2017 Downloaded from http://pubs.acs.org on July 19, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of the American Chemical Society is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Carbon Dioxide in a Supported Ionic Liquid Membrane: Structural and Rotational Dynamics Measured with 2D IR and Pump-Probe Experiments Jae Yoon Shin, Steven A. Yamada, and Michael D. Fayer* Department of Chemistry Stanford University, Stanford, CA 94305 Phone: (650)723-4446; Email: *[email protected] Abstract Supported ionic liquid membranes (SILMs) are porous membranes impregnated with ionic liquids (ILs) and used as advanced carbon capture materials. Here, two-dimensional infrared (2D IR) and IR polarization selective pump-probe (PSPP) spectroscopies were used to investigate CO2 reorientation and spectral diffusion dynamics in SILMs. The SILM contained 1ethyl-3-methylimidazolium bis(trifluoromethylsulfonly)imide in the poly(ether sulfone) membrane with average pore size of ~350 nm. Two ensembles of CO2 were observed in the SILM, one in the IL phase in the membrane pores and the other in the supporting membrane polymer. CO2 in the polymer displayed a red-shifted IR absorption spectrum and a shorter vibrational lifetime of the asymmetric stretch mode compared to the IL phase. Despite the relatively large pore sizes, the complete orientational randomization of CO2 and structural fluctuations of the IL (spectral diffusion) in the pores are slower than in the bulk IL by ~2-fold. The implication is that the IL structural change induced by the polymer interface can propagate out from the interface more than a hundred nanometers, influencing the dynamics. The dynamics in the polymer are even slower. This study demonstrates that there are significant differences in the dynamics of ILs in SILMs on a molecular level compared to the bulk IL, and the study of dynamics in SILMs can provide important information for the design of SILMs for CO2 capture.

1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 44

I. Introduction The rise in the atmospheric concentration of the major anthropogenic greenhouse gas, carbon dioxide, has been accelerated since the industrial revolution, resulting in the onset of global warming and climate change that will cause severe effects on humanity.1-3 Thus, reducing the current level of CO2 emission into the atmosphere is important to global society, and carbon capture and storage (CCS) from stationary sources is an important goal.4-5 In the CCS process, CO2 is captured before it enters to the atmosphere. Currently, the capture process is expensive, inhibiting the practical application of CCS. In many approaches, regenerating the capture material after CO2 separation requires substantial energy.5 Therefore, developing efficient methods and materials for CO2 capture is the key to the accomplishments of CCS. Ionic liquids (ILs) are molten salts, consisting of complex cations and anions, that have melting points below 100 °C.6 ILs that remain as liquids at room temperature are called RTILs. RTILs have been of great interest as one of promising media for carbon capture because they have negligible volatility,7 low flammability,8 high thermal stability,9 and more importantly, high tunability of their physical and chemical properties.6, 10 Degradations and loss of the capture medium, which are major flaws of using organic solvents in carbon capture, can be essentially eliminated by using RTILs.11-13 Supported ionic liquid membranes (SILMs) are advanced materials for carbon capture and separation. The SILMs are, in general, based on organic or inorganic porous membranes whose pores are filled with RTILs.14-17 SILMs are highly cost and energy efficient because they have a large surface area for CO2 capture and diffusion of CO2 through the RTIL in the pores provides an avenue for capture of the CO2 without the need for a solvent regeneration process.1417

Therefore, various types of SILMs have been developed and examined to optimize their

2


Page 3 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


permeability and selectivity for applications.17 In addition, there has been extensive studies on nanoconfined ILs in various porous materials to understand physicochemical properties and dynamics of ILs on a molecular level, using different experimental techniques.18-23 Yet, ultrafast spectroscopic studies of these systems are rare although the fundamental time scales of molecular dynamics are usually on picosecond scale. Recently, ultrafast time-resolved IR spectroscopies were employed to investigate the dynamics of CO2 in bulk RTILs.24-26 These bulk IL experiments were conducted as a function of alkyl chain length on 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (EmimNTf2) and with 4, 6, and 10 carbon chain lengths on the imidizolium. Prior to the study of CO2 in these bulk RTILs, the small anion, SeCN–, was used as the vibrational probe to investigate the same series of bulk RTILs.27-28 In the first study of the dynamics of RTILs in SLIMs, SeCN in EmimNTf2 was used as the vibrational probe. The results of these studies using ultrafast two dimensional infrared (2D IR) spectroscopy and IR polarization selective pump-probe (PSPP) experiments demonstrated that the RTIL dynamics in the SILMs was substantially slower than in the bulk.29 Therefore, the dynamical properties of ILs in SILMs cannot be obtained from studies of the bulk liquid.29 Thus, measurements of molecular interactions and dynamics in SILMs are important for developing an understanding of SILMs for technological applications. In this study, the orientational and structural dynamics were investigated using PSPP and 2D IR spectroscopies applied to SILMs containing 13CO2, which were prepared with EmimNTf2 and poly(ether sulfone) (PES) membranes. Although the same SILM was previously examined with SeCN– vibrational probe due to the ease of sample preparation,29 CO2 is the relevant solute for investigating SILMs for CO2 capture. Unlike SeCN–, 13CO2 in the SILM samples existed in

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 44

two distinct ensembles, that is, in the IL and in the polymer itself. Two well resolved CO2 asymmetric stretch bands were observed in the Fourier transform infrared (FT-IR) spectrum of the SILM sample. The blue shifted (higher frequency) band is the same as observed in the bulk RTIL. The CO2 observed in the additional (lower frequency) band in the membrane polymer exhibits very different behavior from the ensemble in the IL phase. The dynamics of the CO2 ensemble in the polymer are much slower than they are in the IL in the SILM, which in turn are slower than the dynamics in the bulk IL. The PES membrane used in this study has very large pores, with the average pore size of ~350 nm.29 The fact that both the CO2 and SeCN vibrational probes report on substantial slowing of the RTIL dynamics in such large pores is remarkable. For example, water in nanoscopic confinement in systems, like AOT reverse micelles, has bulk-like water behavior away from the interface once the diameter is greater than ~5 nm.30-31 Moreover, pulsed field gradient NMR study found that the diffusivities of CO2 and IL decrease under condition of nanoconfinement in KIT-6 silica that has pore size of 8.5 nm, compared to the bulk condition.19 The results presented here show that pore interface in SILMs produces long-range IL structuring that is different from the bulk IL structure. The change in IL structure in the pores and the slowing of the observed dynamics suggests that CO2 translational diffusion will also be slowed. The translational diffusion coefficient of CO2 in the pores, which is directly connected to the SILM’s performance, was estimated using the complete orientational relaxation time measured by PSPP and the Stokes–Einstein (SE) equation.32-33 II. EXPERIMENTAL PROCEDURES A. Sample Preparation

4


Page 5 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


EmimNTf2 was purchased from Iolitec and stored in a nitrogen glove box after drying under vacuum (~ 100 mTorr) at ~65 °C. Isotopically labeled 13CO2 ( 50 ps for both 13CO2 in the IL and in the polymer are extended back to t = 0, the curves are above the short time data. This behavior has

12


Page 13 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


been observed previously for 13CO2 in bulk EmimNTF2.24 As shown in Figure 1A, to the red of the main bands is a hot band that is the result of the thermal population of the bending modes. When a bending mode is thermally populated, the frequency of the stretch is shifted to the red (combination band shift). The bandwidth of the IR pulses is substantially larger than the difference in frequency between the stretch and the combination band, so the excitation pulses populates both modes. When the thermally excited bend relaxes to the ground state, population is transferred from the combination band to the stretch band. When a bend becomes thermally excited, population is transferred from the stretch band to the combination band. However, the combination band has a larger transition dipole than the stretch.24 Therefore, the small combination band is over populated relative to the thermal equilibrium ratio of populations, and there will be a net flow of population from the combination band to the stretch band. After several bend lifetimes, the excited state populations of the two bands will be in their thermal equilibrium ratio. After that, population exchange will continue, but it will not influence the long time (>50 ps) exponential lifetime decay of the stretch band. In addition to influencing the population decays, population exchange between the hot band and the stretch band produces off-diagonal peaks that grow in with the bend lifetime.24 This process has been discussed in detail for 13CO2 in the bulk IL.24 The only difference here is that there are two sets of off-diagonal peaks, one set involving the 13CO2 in the IL in the membrane pores and the other set involving the 13CO2 in the membrane polymer. Within experimental error, the bend lifetimes in both membrane environments are the same as found in the bulk IL, 13 ps. This very low amplitudes off-diagonal peaks do not influence the analysis of the diagonal peaks to give the time dependence of the spectral diffusion and the FFCF.

13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 44

Within the experimental error, the 13CO2 asymmetric stretching mode lifetime is the same in the IL in the membrane pores as it is in the bulk IL. The lack of change in the vibrational lifetime in the SILM pores was also observed in the study of the same SILM with SeCN– as the vibrational probe.29 As discussed briefly above, vibration relaxation is very sensitive to the local environment. The fact that the vibrational lifetimes are the same in the bulk and in the pores for two vibrational probes indicates that there is no substantial change in the structure of the IL surrounding the probes. C. PSPP Measurements: Reorientation Dynamics

The anisotropy data from the PSPP measurements were utilized to determine the orientational relaxation of the 13CO2 (see Equation 4). Figure 3 presents the anisotropy decays of 13

CO2 in the bulk EmimNTf2, in the IL in the SILM pores, and in the membrane polymer. The

data were taken at the center frequency of each of the asymmetric stretch bands. All of the anisotropy decays in Figure 3 begin with initial values less than the maximum anisotropy value, 0.4, when the data are extrapolated back to t = 0. This initial drop from 0.4 is caused by ultrafast inertial orientational motion that occurs on a faster time scale than the time duration of the overlapped pulses and cannot be resolved.43 The difference between 0.4 and the t = 0 value gives information on the extent of inertial orientational motion. After the initial drop, the anisotropy curves are fit well by triexponential decays (Figure 3), and their fit parameters are summarized in Table 1. No offsets are needed in the fits of the anisotropy decays for the 13CO2 in EmimNTf2 in the SILM pores and the bulk sample, indicating that there is complete orientational relaxation. In contrast, for 13CO2 in the polymer of the SILM, an offset is necessary in their fits because the time range in which the data can be collected is limited by the vibrational lifetime and is not long enough to determine the complete anisotropy curve. It is difficult to determine the offset level in

14


Page 15 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the fit, which results in a large error in the longest time constant, t3. If the offset is fixed to 0 in the fit, the first two time constants do not change, but t3 becomes very long, well outside the experimental range but has a smaller error because the long time constant and the offset don’t play off against each other. As will be discussed below, the 2D IR experiments suggest that the 13

CO2 ensemble in the polymer is unlikely to undergo complete orientational randomization. The

2D IR data led to fitting the anisotropy data from the membrane phase with a triexponential decay with an offset. With either fitting procedure, it is clear that the final complete orientation relaxation either does not occur or it is exceedingly slow. As depicted in Figure 3, the anisotropy decays of 13CO2 slows down in the IL in the pores and in the polymer. In going from the bulk sample to the IL phase in the SILM, the longest time constant, t3, changes from 51 to 90 ps while the first two time constants, t1 and t2, remain unaffected within experimental error. The triexponential anisotropy decays in EmimNTf2 for both the bulk and in the IL in the SILM samples, after the inertial drop, can be interpreted as two sequentially restricted angular diffusions (wobbling-in-a-cone) followed by complete orientational randomization of the 13

CO2.44-47 Using the standard approach (wobbling analysis),44-45, 48-49 the cone angles are

retrieved from the amplitudes of the exponential terms in the anisotropy decay, and the correlation times are given by τc1 = t1 for the first cone, τc2 = (1/t2 – 1/t3)–1 for the second cone, and τm = t3 for the complete orientational randomization. The orientational diffusion constants for each cone are calculated based on both cone angle and correlation time.44 The inertial cone angle is recovered from the extent of the initial drop, but its time dependence cannot be resolved. The resulting parameters from the wobbling analysis are tabulated in Table 2. The inertial cone angle of the IL phase in the SILM sample is somewhat smaller than that in the bulk sample. The local 15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 44

IL structure in the SILM allows 13CO2 to have smaller space available for ballistic inertial motion than in the bulk sample. This is an indication that the IL structure surrounding the 13CO2 is not identical to the bulk liquid, in contrast to the spectra and lifetimes, which were not changed within experimental error. The second cone angle in the pores is also slightly smaller compared to the bulk sample whereas the first cone angles in the both samples are almost the same. The decrease in both inertial and second cone angles of the SILM sample leads to the decrease in the total cone angle that is sampled prior to the complete orientational randomization. However, the overall differences in the cone angles between the bulk and SILM samples are small; there is only subtle change in the local environment experienced by 13CO2. Within the error bars, the time constants for the first and second cones are very close for the two samples, and thus, given the small changes in the cone angles, the diffusion constants are similar. The similarities in the cone angles, correlation times, and diffusion constants reflect that the local environments around 13

CO2 are very similar in the pores and bulk sample. This is consistent with the previous results,

using SeCN as the probe, that the local IL structure is essentially unaffected by confinement in the pores of the PES200 membrane.29 Unlike the wobbling-in-a-cone, the time constant for the complete orientational randomization, τm, increases from 51 to 90 ps in going from the bulk to the SILM sample, revealing significant slowdown. (The diffusion constant is Dm = 1/6τm.) In the previous study, it was found that τm of 13CO2 in the bulk EmimNTf2 agrees well with the orientational diffusion time constant calculated from the Debye–Stokes–Einstein (DSE) equation under the slip boundary conditions.25 In addition, τm increased in proportion to the viscosity of the bulk IL as the alkyl chain of the cation of the IL became longer.26 This suggests that the rearrangement of global IL structure is necessary for CO2 to undergo the complete orientational randomization and 16


Page 17 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


for the bulk ions to move out of the way. Therefore, the increase of τm in the SILM sample suggests that the global IL structure is substantially influenced by the confinement in the membrane pores. The confinement effects on the IL in the PES200 membrane as shown through the deviation from the bulk liquid dynamics were unexpected because the pore size is very large compared to the sizes of other systems where nanoconfinement strongly affects liquid dynamics.30-31, 50-52 The IL structure at surfaces and interfaces has been of great interest recently, and it is commonly accepted that the interactions with interfaces can affect the ordering of cations and anions within a few nanometers from interfaces.53-59 However, this picture has been challenged by recent experimental and theoretical studies which indicate that ILs can form ordered structures distinct from the bulk over much large distances from an interface.60-63 In line with these recent studies, our results provide substantial evidence that the poly(ether sulfone) interface can influence the global IL structure over a large distance. The slowdown of complete orientational relaxation in the SILM was also observed with SeCN– as the vibrational probe. In going from the bulk EmimNTf2 to the same SILM studied here, τm of SeCN– increased from 136 to 512 ps.29 This 3.8-fold increase is substantially larger than 1.8-fold increase observed here for CO2. The similar trend also can be found in alkyl chain length dependence of τm. As the IL changed from EmimNTf2 to DmimNTf2 (1-decyl-3methylimidazolium bis(trifluoromethylsulfonyl)imide), τm increased by a factor of 3.6 for SeCN– and 1.8 for CO2.26-27 The carbon dioxide molecule is relatively small and is known to occupy empty cavities in the IL.64 In addition, CO2 is a neutral molecule, mainly interacting with anions in the IL65-66 whereas SeCN– is an anion, probably interacting more strongly with cations in the IL.

17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 44

To qualitatively address the extent of the slowing of the IL rearrangement experienced by CO2 in the SILM, an “effective viscosity” can be determined. With the reasonable assumption that the shape and volume of CO2 does not change when confined in the SILM, τm only depends on the viscosity according to the DSE equation. However, the alkyl chain length dependent experiments with the 13CO2 vibrational probe found that τm does not perfectly follow the bulk viscosity changes.26 To compensate for this difference, a correlation plot of τm measured in different IL versus the bulk viscosity of each IL was fit with linear functional form to obtain the empirical relationship. Then, the effective viscosity experienced by CO2 in the SILM can be obtained using τm of the SILM sample and the empirical functional form. The calculated effective viscosity in the SILM is 117.0 cP which is close to the bulk viscosity of DmimNTf2 (134.4 cP).26 The viscosity of bulk EmimNTf2 is 36.3 cP. The effective viscosity of the IL in the pores is ~3 times greater than the bulk viscosity. This increase in the viscosity is quite remarkable because CO2 is in the same IL in both bulk and SILM samples and only difference is the presence of polymer interface in the SILM sample. The long-range IL ordering effect caused by the pore interfaces substantially influences the global IL dynamics. D. 2D IR Measurements: Spectral Diffusion and Structural Dynamics

2D IR spectroscopy measures structural evolution of a system by reporting on the spectral diffusion of a vibrational probe that senses the dynamics of surrounding liquid structures. The vibrational probe molecules will experience different interactions with their surroundings because of the inhomogeneity in the liquid structures, giving rise to the inhomogeneous broadening of the vibrational absorption line. Spectral diffusion provides information on how long it takes the probe molecules to sample all the frequencies in the inhomogeneously broadened line, i.e., all the liquid structures that contribute to inhomogeneous

18


Page 19 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


broadening of the absorption spectrum. The amplitudes and time scales of the spectral diffusion can be quantified through the FFCF, which is the probability that a vibrational oscillator with an initial frequency in the inhomogeneous spectral distribution will have that same frequency at a later time, averaged over all initial frequencies. The FFCF is modeled with a multiexponential form

C (t )   (t ) (0)   i2 exp(t /  i )

(5)

i

where ∆i and τi are the frequency fluctuation amplitudes and associated time constant, respectively, for the ith component. (t )  (t )   is the instantaneous frequency fluctuation, where  is the average frequency. A component of the FFCF is motionally narrowed if   1 , and then its ∆ and τ cannot be determined separately. The motionally narrowed component contributes to the homogeneous broadening of the absorption line and has a pure dephasing linewidth, given by *  2 /   1/ ( T2* ) , where T2* is the pure dephasing time. The vibrational lifetime and orientational relaxation also contribute to the homogeneous broadening through the total homogeneous dephasing time, T2 which is given by

1 1 1 1  *  T2 T2 2T1 3Tor

(6)

where T1 and Tor are the vibrational lifetime and orientational relaxation time, respectively. Then, the total homogeneous linewidth is Γ = 1/(πT2). As mentioned in the Experimental Procedures, the FFCF can be extracted from the line shape analysis of 2D IR spectra, using the CLS method.38-39 Figure 4 presents 2D IR spectra of 13CO2 in the SILM sample at Tw = 4 and 60 ps. Two diagonal bands are from two ensembles of 13CO2 residing in the IL and the polymer that makes 19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 44

up the membrane. At a short waiting time, Tw = 4 ps (Figure 4A), both bands are elongated along the diagonal (dashed white line) because the environment has had little time to evolve and the vibrational oscillators have frequencies close to their initial frequencies. However, the extent of elongation for each band is different, with 13CO2 in the polymer having a higher degree of elongation than that for 13CO2 in the IL in the membrane pores. At a longer waiting time, Tw = 60 ps (Figure 4B), the shape of IL phase band becomes more round due to the evolution of IL structures. In contrast, the band shape of the supporting membrane phase is not very different from the initially elongated shape at Tw = 4 ps, indicating its slow spectral diffusion. Note that the spectra are normalized to the highest amplitude contour. In Figure 4B, the IL band has become relatively higher in amplitude because it has a longer lifetime. At very long time, the IL band is larger in amplitude than the polymer band. As discussed in the Experimental Procedures, 2D IR experiments were performed in

 XXXX  (parallel) and  XXYY  (perpendicular) polarization configurations. It was recently observed that the CLS decays obtained with these two polarization pathways are not necessarily the same, demonstrating that, in certain cases, structural fluctuations of the sample medium do not entirely determine the frequency fluctuations of the vibrational probe.67 The rotation of the molecule can also play a role. The difference in the parallel and perpendicular decay curves was addressed by modeling the interaction of the vibration with its surroundings as coupling to electric fields, Stark effect, produced by the liquid in which the vibrational probe is embedded. For CO2, the coupling is through the second order Stark effect.25, 67-68 When the field varies on a similar timescale, or slowly, relative to the orientational relaxation times of the probe molecule, frequency fluctuations occur due to the probe rotation. Therefore, reorientation of the probe molecule contributes to spectral diffusion and gives rise to different CLS decays for  XXXX  20


Page 21 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and  XXYY  polarization configurations.25, 67-68 This phenomenon is referred to as reorientation-induced spectral diffusion (RISD), in contrast to structural spectral diffusion (SSD), which arises solely from structural evolution of the liquid. In general, we are interested in measuring SSD. Analytical expressions are available for the RISD contribution. The expressions use as inputs the orientational relaxation measured with PSPP experiments as their inputs. Therefore, the SSD can be obtained from fitting the parallel and perpendicular 2D IR data. The second-order Stark effect RISD theory for CO2 results a further separation of SSD into the contributions from scalar and vector interactions of the medium with CO2. A more detailed discussion including the equations used in the data fitting are presented in Supporting Information. Figure 5A shows 2D IR CLS data (points) for both the bulk sample and the IL in the membrane pores in the XXXX and XXYY polarization configurations. The data for each sample are simultaneously fit with the second-order Stark model described briefly above. The scalar and vector correlation functions were modeled with the same functional forms as previously used for the bulk sample:25-26 a biexponential for the scalar part and a single exponential for the vector part. The model fits the data very well, as can be seen in Figure 5A (solid lines). If single exponentials are used for each term, the model does not fit the data. In addition, adding more exponentials does not improve the fit quality and does not result in unique time constants. In the SILM sample, XXXX CLS curve of IL phase band (2277 cm-1) showed a small oscillatory feature starting with growth at short waiting time from 0.3 ps up to 3 ps. The same feature starting with decay also appeared in its XXYY CLS. (The features are not shown here.) The oscillations in CLS curves have previously been observed in D2O in RTIL. The coherent 21



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 44

energy transfer between symmetric and antisymmetric stretch modes of D2O can cause the oscillations in the 2D line shapes.69 The oscillatory feature observed in the SILM sample could also indicate the presence of the coherent excitation transfer, presumably between asymmetric stretch mode of CO2 and its hot band; the beating pattern appears differently in two polarization configurations. Another possibility is that there is a fast rapidly damped physical oscillation of the CO2 relative to the cations it interacts with. This type of oscillation has been observed HOD in D2O at very short times (~100 fs)70 and on a longer time scale (1.5 ps) for benzonitrile in a metal organic framework.71 As the oscillation is not observed in the bulk sample, the second explanation may be correct and indicate a difference in the CO2 – IL interaction in the membrane pores compared to the bulk IL. However, the exact origin is unclear. For this reason, the early time CLS points up to 4 ps were excluded in the RISD fit for the SILM sample. Since the time scales of CLS decays are significantly longer than that of early time features, the results from the RISD fit will not be affected. Table 3 summarizes the SSD parameters obtained from the secondorder Stark model fits, comparing CO2 in the bulk EmimNTf2 and in the IL phase of SILM sample. As can be seen in Figure 5A, the CLS decays of CO2 in the IL phase are substantially slower in the SILM sample. In going from the bulk to SILM samples, the longer time constant, τ2, in the scalar correlation function and the vector correlation time, τ3, increase by a factor of 2 while the shorter time constant, τ1, in the scalar part remains the same as found for the bulk, within the error bars. The similar changes in the τ2 and τ3time scales in going from the bulk IL to IL in the pores may suggest that both time constants originate from the same IL motions that can alter both the magnitude and direction of the solvent’s internal electric fields. In contrast, the short scalar correlation time, τ1, is unchanged in going from the bulk to the pores. This

22


Page 23 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


component was previously observed to display no dependence on the alkyl chain length of bulk RTILs, which suggests its insensitivity to the extent of nonpolar aggregation of the IL and thus, the overall IL structures.26 It may involve very local IL motions that are not influenced by either alkyl chain length or confinement in the membrane pores. In contrast, it was observed that the two SSD components, τ2 and τ3, are sensitive to the global structures and overall fluctuations of IL.26 In the CO2 study in the bulk RTILs, τ2 and τ3 became longer with increasing alkyl chain length of IL. The chain length dependence in the bulk indicates that the slowdown in τ2 and τ3 in the SILM sample reflects a change in the global IL structure and their dynamics. The slowing of the spectral diffusion and the slowing of the orientational relaxation in the SILM relative to the same IL in the bulk demonstrates that the IL dynamics, and therefore structure, is significantly influenced by the presence of the pore interface, in spite of the large pore sizes. The same trend in SSD was observed for the SeCN– in SILM.29 Although the polarization dependent CLS data allows us to determine the SSD part of the total FFCF without the RISD contributions, MD simulations will determine the complete FFCF without a separation into SSD and RISD components.72 Therefore, the isotropic decay is obtained by taking the parallel 2D spectrum and adding twice the perpendicular spectrum with the parallel and perpendicular spectra properly normalized. The resulting isotropic curve, which still contains the RISD contribution, is useful because it is the CLS that would be calculated in a molecular dynamics (MD) simulations, which in general, do not include the polarized radiation fields. The isotropic FFCF parameters obtained from the isotropic CLS decays are given in Table 4.

23



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 44

In contrast to the IL CLS data of the SILM and the bulk, the membrane phase data did not yield data that is consistent with the RISD theory. The parallel and perpendicular measurements on 13CO2 in the polymer did not show the characteristic behavior in which the perpendicular decay is faster than the parallel decay. The RISD theory was derived for vibrational probes that are orientationally distributed randomly relative to the electric fields produced by the surrounding medium. If the CO2 enters the polymer strands with an orientation that is correlated with the strand topography and it does not undergo complete rotational relaxation, then it is likely that the CO2 will not be randomly oriented relative to the polymer produced electric field. This is a situation that is not covered by the RISD theory. As discussed in connection with the orientational relaxation of CO2 in the polymer, there is a very slow component. The slow component is either an offset, which means that complete orientational relaxation does not occur, or it is a very slow component. However, a very slow component, even if it is much slower than the vibrational lifetime, will still result in random orientations of the CO2 relative to the polymer, in which case the RISD theory should work. Vibrational probes that are coupled to electric fields by either first or second order Stark effects have displayed the features described by RISD theory for a wide variety of systems.25-28, 67-68 Therefore, it is likely that the CO2 does not undergo complete orientational relaxation in the polymer. While RISD theory cannot be applied to separate SSD and RISD, we can still compare the spectral diffusion of CO2 in the polymer vs. the bulk IL and the IL in the membrane pores. Figure 5B shows the perpendicular data sets for the three environments, CO2 in the bulk IL, in the IL in the pores, and in the polymer. The structural dynamics in the polymer is vastly slower than in either the bulk IL or the IL in the pores. As the main interest is the difference in the

24


Page 25 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


dynamics in the IL in the pores vs. in the bulk IL, no further analysis of dynamics of CO2 in the polymer will be given. Comparison of the CO2 SSD for the bulk and IL in the pores shows that spectral diffusion in the pores is substantially slower than in the bulk. The comparison of the orientational relaxation displays the same behavior. The interfaces of pores induce the reconfigurations of IL’s global structures, presumably through the long-range ordering effect, slowing down the dynamics. In the membrane phase of SILM, the spectral diffusion of CO2 is far slower due to the direct interactions with the polymer. E. Estimate of Translational Diffusion in the Pores

For practical application as a carbon capture material, the ability of CO2 to diffuse through the SILM, that is diffusivity, is one of important factors together with solubility.17 A estimate of the translational diffusion in the pores can be made using the effective viscosity given above. If the molecular motions are reasonably hydrodynamic, both rotational and translational diffusion will be governed by the same friction term. In cases where the solute molecule is relatively small compared to the solvent, like CO2 in the IL, then both rotational and translational diffusion will have slip boundary conditions rather than stick. The standard form of the Stokes–Einstein (SE) equation is for stick boundary conditions and spherical particles. With slip boundary conditions, diffusion occurs more rapidly than predicted by the standard SE equation.32 The standard SE equation for diffusion can be modified for a small and nonspherical molecule by including a shape factor and size factors in the equation.32-33 A modified SE equation gives the translational diffusion constant as DT 

k BT , c r f

(7)

25



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 44

where kB is Boltzmann’s constant, T is absolute temperature, and η is the viscosity.

r  3 / (4 ) Veff1/3 is the effective radius, where Veff is the molecular volume. The size factor, c, accounts the relative hydrodynamic dimensions of the diffusing species and solvent and can be derived from the microfriction theory.73 The shape factor, f, is used for the nonspherical diffusing particles and has been theoretically calculated for prolate and oblate ellipsoids.74 If we assume that none of the molecular properties change in going from the bulk IL to the IL in the pores, i.e., the molecular dimensions, the CO2 shape factor, then the diffusion constant in the IL pores is determined by the ratio of the viscosity for the bulk IL (36.3 cP) and the effective viscosity determined from the orientation relaxation time for the IL in the pores (117 cP). The experimental value in the bulk IL is 5 × 10-6 cm2s-1.75-76 Then, the estimated value for the IL in the SILM pores is 1.6 × 10-6 cm2s-1. Although there is no experimentally determined diffusion constant that can be directly compared, the diffusivity can be extracted from the permeability of CO2 measured in the same SILM. Scovazzo et al. have measured the CO2 permeability through the SILM prepared with PES200 and EmimNTf2, using the time-lag method in gas flux experiments.77 Because the permeability is a product of solubility and diffusivity according to the solution-diffusion model,17 the diffusivity can be calculated based on the reported permeability (960 barrers) and CO2 gas solubility in the bulk EmimNTf2 (0.1 mol/L atm).77 It should be noted that the CO2 solubility used in this calculation is for the bulk EmimNTf2, not for the SILM. The value for the SILM is unavailable in the literature. The calculated diffusivity, 3.2 × 10-6 cm2s-1, is larger than the diffusion constant obtained from the effective viscosity determine using the orientational relaxation measurements by a factor of 2. In the permeability experiments, a pressure gradient is applied across the SILM to drive the diffusion of CO2 in the gas flux measurements.32 The 26


Page 27 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


presence of driving force can accelerate the diffusion of CO2, which results in larger diffusion constant. Another possibility is that the CO2 solubility in the SILM might be larger than that in the bulk IL. A recent simulation study demonstrated that nanoconfinement of IL in model membranes with very small 2 and 5 nm pores increase CO2 solubility compared to the bulk IL.78 Although the pore sizes in the simulation are much smaller than the SILM used in this study, the solubility could still be affected by the confinement in the SILM as the CO2/IL dynamics are influenced. Nonetheless, both the permeability measurements and the orientational relaxation measurements indicate that the translational diffusion in the IL in the pores is slower than in the bulk IL, presumably because of the long-range IL ordering caused by the pore interface. IV. Concluding Remarks

The dynamics of 13CO2 in a SILM prepared with EmimNTf2 and PES200 polymer membranes were investigated by the polarization selective pump-probe and 2D IR spectroscopies and compared to the corresponding bulk sample. The FT-IR spectrum of 13CO2 in the SILM sample showed a bulk-like absorption bands and an additional band that is red-shifted by 6 cm-1 from the bulk IR absorption spectrum. Since the red-shifted band was observed in the 13

CO2 samples in PES200 with and without the IL in the pores (Figure 1), the additional features

in the SILM sample can be assigned to the CO2 in the polymer that makes up the membrane. The population data analysis found that the lifetimes of asymmetric stretch mode are the same in the bulk IL and in the IL in the SILM pores but the lifetime is shorter in the polymer. The facts that the 13CO2 spectra and the lifetimes in the bulk IL and the IL in the pores are the same, demonstrates that there is no major difference in the very local environment surrounding the 13

CO2.

27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 44

The orientational relaxation anisotropy decays showed substantial differences between the bulk IL and IL in the membrane samples. For the CO2 in the IL phase in the bulk and SILM samples, the anisotropy decays were analyzed by the wobbling-in-a-cone model. While the half angles and time constants for two wobbling cones were similar in the bulk and pores, the complete orientational relaxation time increased by a factor of ~2 in going from bulk to SILM. The effective viscosity that CO2 experiences in EmimNTf2 confined in the SILM pore was found to be 117.0 cP compared to the bulk viscosity of 36.3 cP. The translational diffusion constant of CO2 in SILM was found approximately using the effective viscosity. The value of 1.6 × 10-6 cm2s-1 is approximately a factor of three smaller than the value in the bulk. The slowing of the dynamics in SILM was also demonstrated by polarization selective 2D IR experiments. Like the anisotropy decays, the fastest components of the spectral diffusion in the bulk IL and in the IL in the pores are the same within error. However, the slower two components are slower in the pores than in the bulk by a factor of ~2. Both the anisotropy data and the 2D IR data taken on the absorption band corresponding to CO2 in the polymer show dramatic slowing compared to CO2 in the IL. Given that the polymer is a solid, this is not surprising. Experiments that cannot discriminate between CO2 in the IL in the pores and in the polymer may need to consider that there are two distinct subensembles of CO2 in the membrane that have very different dynamics. CO2 in the PES200 SILMs like a previous study of SeCN‒ in the same SILM29 shows that the influence of confinement in the pores is significant. The pores in these membranes have an average diameter of ~350 nm.29 This diameter is large, approaching macroscopic size. Other types of liquids in pores of this size would not show confinement effects. Only molecules very near the interface would be effected. These experiments and other experiments53-54, 56-57, 59-62 and

28


Page 29 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


simulations55, 58, 63 show that the interactions of an IL with an interface are very different from those of other liquids. ILs in the bulk always show charge ordering. The influence of the interface may be to produce charge ordering at the interface that is very different from that found in the bulk. Then strong Coulomb interactions among the ions apparently propagate this structural difference over long distances. Understanding how interfaces influence the structure and dynamics of ILs is an interesting and important topic. Finally, the main purpose of SILMs is to capture CO2. For practical applications, CO2 selectivity and transport rate are important for the performance of a SILM. Then, the properties of both RTILs and supporting membranes must be tested and optimized. The results presented here demonstrate that the dynamics of ILs confined in SILMs are different from those found in the bulk IL. Long-range ordering effect near the interfaces can slow down the structural dynamics of IL and, therefore, the orientational and translational diffusions of CO2. The properties and dynamics of ILs in SILMs cannot necessarily be predicted from the bulk IL. Then, fundamental understanding of molecular interactions of RTILs and their dynamics in SILMs is important for developing SILMs for application in CO2 capture.

Supporting Information

Details of the second-order Stark effect RISD theory.

Acknowledgements

This work was funded by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences of the U.S. Department of Energy through Grant

29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 44

#DE-FG03-84ER13251. Additional support came from the Stanford Center for Carbon Storage, which provided partial salary support for J.Y.S.

30


Page 31 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References

1.

Dlugokencky, E.; Tans, P. NOAA/ESRL: http://www.esrl.noaa.gov/gmd/ccgg/trends/

2.

Solomon, S.; Plattner, G.-K.; Knutti, R.; Friedlingstein, P., Proc. Natl. Acad. Sci. U. S. A.

2009, 106, 1704-1709.

3.

Shakun, J. D.; Clark, P. U.; He, F.; Marcott, S. A.; Mix, A. C.; Liu, Z.; Otto-Bliesner, B.;

Schmittner, A.; Bard, E., Nature 2012, 484, 49-54. 4.

Metz, B.; Davidson, O.; Coninck, H. d.; Loos, M.; Meyer, L., IPCC Special Report on

Carbon Dioxide Capture and Storage. Prepared by Working Group III of the Intergovernmental Panel on Climate Change. Cambridge University Press: Cambridge, United Kingdom and New York, NY, USA, 2005; p 442. 5.

Haszeldine, R. S., Science 2009, 325, 1647-1652.

6.

Plechkova, N. V.; Seddon, K. R., Chem. Soc. Rev. 2008, 37, 123-150.

7.

Earle, M. J.; Esperanca, J. M. S. S.; Gilea, M. A.; Canongia Lopes, J. N.; Rebelo, L. P.

N.; Magee, J. W.; Seddon, K. R.; Widegren, J. A., Nature 2006, 439, 831-834. 8.

Smiglak, M.; Reichert, W. M.; Holbrey, J. D.; Wilkes, J. S.; Sun, L.; Thrasher, J. S.;

Kirichenko, K.; Singh, S.; Katritzky, A. R.; Rogers, R. D., Chem. Commun. 2006, 2554-2556. 9.

Anderson, J. L.; Ding, R.; Ellern, A.; Armstrong, D. W., J. Am. Chem. Soc. 2005, 127,

593-604. 10.

Castner, E. W.; Margulis, C. J.; Maroncelli, M.; Wishart, J. F., Annu. Rev. Phys. Chem.

2011, 62, 85-105.

11.

MacDowell, N.; Florin, N.; Buchard, A.; Hallett, J.; Galindo, A.; Jackson, G.; Adjiman,

C. S.; Williams, C. K.; Shah, N.; Fennell, P., Energy Environ. Sci. 2010, 3, 1645-1669.

31



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

12.

Page 32 of 44

Kenarsari, S. D.; Yang, D.; Jiang, G.; Zhang, S.; Wang, J.; Russell, A. G.; Wei, Q.; Fan,

M., RSC Adv. 2013, 3, 22739-22773. 13.

Boot-Handford, M. E.; Abanades, J. C.; Anthony, E. J.; Blunt, M. J.; Brandani, S.; Mac

Dowell, N.; Fernandez, J. R.; Ferrari, M.-C.; Gross, R.; Hallett, J. P.; Haszeldine, R. S.; Heptonstall, P.; Lyngfelt, A.; Makuch, Z.; Mangano, E.; Porter, R. T. J.; Pourkashanian, M.; Rochelle, G. T.; Shah, N.; Yao, J. G.; Fennell, P. S., Energy Environ. Sci. 2014, 7, 130-189. 14.

Scovazzo, P.; Visser, A. E.; Davis, J. H.; Rogers, R. D.; Koval, C. A.; DuBois, D. L.;

Noble, R. D., Supported Ionic Liquid Membranes and Facilitated Ionic Liquid Membranes. In Ionic Liquids, American Chemical Society: Washington, DC, 2002; Vol. 818, pp 69-87. 15.

Lozano, L. J.; Godínez, C.; de los Ríos, A. P.; Hernández-Fernández, F. J.; Sánchez-

Segado, S.; Alguacil, F. J., J. Membr. Sci. 2011, 376, 1-14. 16.

Dai, Z.; Noble, R. D.; Gin, D. L.; Zhang, X.; Deng, L., J. Membr. Sci. 2016, 497, 1-20.

17.

Tome, L. C.; Marrucho, I. M., Chem. Soc. Rev. 2016, 45, 2785-2824.

18.

Zhang, S.; Zhang, J.; Zhang, Y.; Deng, Y., Chem. Rev. 2017, 117, 6755-6833.

19.

Hazelbaker, E. D.; Guillet-Nicolas, R.; Thommes, M.; Kleitz, F.; Vasenkov, S.,

Microporous Mesoporous Mater. 2015, 206, 177-183. 20.

Iacob, C.; Sangoro, J. R.; Kipnusu, W. K.; Valiullin, R.; Karger, J.; Kremer, F., Soft

Matter 2012, 8, 289-293. 21.

Nayeri, M.; Aronson, M. T.; Bernin, D.; Chmelka, B. F.; Martinelli, A., Soft Matter 2014,

10, 5618-5627. 22.

Iacob, C.; Sangoro, J. R.; Papadopoulos, P.; Schubert, T.; Naumov, S.; Valiullin, R.;

Karger, J.; Kremer, F., Phys. Chem. Chem. Phys. 2010, 12, 13798-13803.

32


Page 33 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


23.

Han, K. S.; Wang, X.; Dai, S.; Hagaman, E. W., J. Phys. Chem. C 2013, 117, 15754-

15762. 24.

Giammanco, C. H.; Kramer, P. L.; Yamada, S. A.; Nishida, J.; Tamimi, A.; Fayer, M. D.,

J. Phys Chem. B 2016, 120, 549-556. 25.

Giammanco, C. H.; Kramer, P. L.; Yamada, S. A.; Nishida, J.; Tamimi, A.; Fayer, M. D.,

J. Chem. Phys. 2016, 144, 104506. 26.

Giammanco, C. H.; Yamada, S. A.; Kramer, P. L.; Tamimi, A.; Fayer, M. D., J. Phys

Chem. B 2016, 120, 6698-6711. 27.

Tamimi, A.; Bailey, H. E.; Fayer, M. D., J. Phys Chem. B 2016, 120, 7488-7501.

28.

Tamimi, A.; Fayer, M. D., J. Phys Chem. B 2016, 120, 5842-5854.

29.

Shin, J. Y.; Yamada, S. A.; Fayer, M. D., J. Am. Chem. Soc. 2017, 139, 311-323.

30.

Fayer, M. D., Physiology 2011, 26, 381-392.

31.

Fayer, M. D.; Levinger, N. E., Annu. Rev. Anal. Chem. 2010, 3, 89-107.

32.

Edward, J. T., J. Chem. Educ. 1970, 47, 261.

33.

Ciancaleoni, G.; Zuccaccia, C.; Zuccaccia, D.; Macchioni, A., NMR Techniques for

Investigating the Supramolecular Structure of Coordination Compounds in Solution. In Techniques in Inorganic Chemistry, CRC Press: 2010; pp 129-180. 34.

Maiella, P. G.; Schoppelrei, J. W.; Brill, T. B., Appl. Spectrosc. 1999, 53, 351-355.

35.

Kumar, S. K.; Tamimi, A.; Fayer, M., J. Chem. Phys. 2012, 137, 184201.

36.

Nishida, J.; Tamimi, A.; Fei, H.; Pullen, S.; Ott, S.; Cohen, S. M.; Fayer, M. D., Proc.

Natl. Acad. Sci. U. S. A. 2014, 111, 18442-18447. 37.

Tokmakoff, A., J. Chem. Phys. 1996, 105, 1-12.

38.

Kwak, K.; Park, S.; Finkelstein, I. J.; Fayer, M. D., J. Chem. Phys. 2007, 127, 124503.

33



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

39.

Kwak, K.; Rosenfeld, D. E.; Fayer, M. D., J. Chem. Phys. 2008, 128, 204505.

40.

Kai, T.; Duan, S., CO2 Capture by Membrane. In Handbook of Climate Change

Page 34 of 44

Mitigation and Adaptation, Chen, W.-Y.; Suzuki, T.; Lackner, M., Eds. Springer New York: New York, NY, 2014; pp 1-28. 41.

Kenkre, V.; Tokmakoff, A.; Fayer, M., J. Chem. Phys. 1994, 101, 10618-10629.

42.

Owrutsky, J.; Raftery, D.; Hochstrasser, R., Annu. Rev. Phys. Chem. 1994, 45, 519-555.

43.

Moilanen, D. E.; Fenn, E. E.; Lin, Y.-S.; Skinner, J.; Bagchi, B.; Fayer, M. D., Proc.

Natl. Acad. Sci. U. S. A. 2008, 105, 5295-5300. 44.

Lipari, G.; Szabo, A., Biophys. J. 1980, 30, 489-506.

45.

Wang, C.; Pecora, R., J. Chem. Phys. 1980, 72, 5333-5340.

46.

Kinosita Jr, K.; Ikegami, A.; Kawato, S., Biophys. J. 1982, 37, 461-464.

47.

Lipari, G.; Szabo, A., J. Am. Chem. Soc. 1982, 104, 4546-4559.

48.

Kramer, P. L.; Giammanco, C. H.; Fayer, M. D., J. Chem. Phys. 2015, 142, 212408.

49.

Tan, H.-S.; Piletic, I. R.; Fayer, M. D., J. Chem. Phys. 2005, 122, 174501.

50.

Richert, R., Annu. Rev. Phys. Chem. 2011, 62, 65-84.

51.

Thompson, W. H., Annu. Rev. Phys. Chem. 2011, 62, 599-619.

52.

Krutyeva, M.; Wischnewski, A.; Monkenbusch, M.; Willner, L.; Maiz, J.; Mijangos, C.;

Arbe, A.; Colmenero, J.; Radulescu, A.; Holderer, O., Phys. Rev. Lett. 2013, 110, 108303. 53.

Antelmi, D. A.; Kékicheff, P.; Richetti, P., Langmuir 1999, 15, 7774-7788.

54.

Mezger, M.; Schröder, H.; Reichert, H.; Schramm, S.; Okasinski, J. S.; Schöder, S.;

Honkimäki, V.; Deutsch, M.; Ocko, B. M.; Ralston, J., Science 2008, 322, 424-428. 55.

Sieffert, N.; Wipff, G., J. Phys. Chem. C 2008, 112, 19590-19603.

56.

Bovio, S.; Podesta, A.; Lenardi, C.; Milani, P., J. Phys Chem. B 2009, 113, 6600-6603.

34


Page 35 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


57.

Mezger, M.; Schramm, S.; Schröder, H.; Reichert, H.; Deutsch, M.; De Souza, E. J.;

Okasinski, J. S.; Ocko, B. M.; Honkimäki, V.; Dosch, H., J. Chem. Phys. 2009, 131, 094701. 58.

Shimizu, K.; Pensado, A.; Malfreyt, P.; Pádua, A. A.; Lopes, J. N. C., Farad. Discuss.

2012, 154, 155-169.

59.

Elbourne, A.; Voïtchovsky, K.; Warr, G. G.; Atkin, R., Chem. Sci. 2015, 6, 527-536.

60.

Parr, D.; Chrestenson, J.; Malik, K.; Molter, M.; Zibart, C.; Egan, B.; Haverhals, L. M.,

ECS Trans. 2015, 66, 35-42. 61.

Jurado, L. A.; Kim, H.; Arcifa, A.; Rossi, A.; Leal, C.; Spencer, N. D.; Espinosa-Marzal,

R. M., Phys. Chem. Chem. Phys. 2015, 17, 13613-13624. 62.

Anaredy, R. S.; Shaw, S. K., Langmuir 2016, 32, 5147-5154.

63.

Amith, W. D.; Hettige, J. J.; Castner, E. W.; Margulis, C. J., J. Phys. Chem. Lett. 2016, 7,

3785-3790. 64.

Huang, X.; Margulis, C. J.; Li, Y.; Berne, B. J., J. Am. Chem. Soc. 2005, 127, 17842-

17851. 65.

Cadena, C.; Anthony, J. L.; Shah, J. K.; Morrow, T. I.; Brennecke, J. F.; Maginn, E. J., J.

Am. Chem. Soc. 2004, 126, 5300-5308. 66.

Hou, Y.; Baltus, R. E., Ind. Eng. Chem. Res. 2007, 46, 8166-8175.

67.

Kramer, P. L.; Nishida, J.; Giammanco, C. H.; Tamimi, A.; Fayer, M. D., J. Chem. Phys.

2015, 142, 184505.

68.

Kramer, P. L.; Nishida, J.; Fayer, M. D., J. Chem. Phys. 2015, 143, 124505.

69.

Wong, D. B.; Giammanco, C. H.; Fenn, E. E.; Fayer, M. D., J. Phys Chem. B 2013, 117,

623-635.

35



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

70.

Page 36 of 44

Fecko, C. J.; Eaves, J. D.; Loparo, J. J.; Tokmakoff, A.; Geissler, P. L., Science 2003,

301, 1698. 71.

Nishida, J.; Fayer, M. D., J. Phys. Chem. C 2017, 121, 11880-11890.

72.

Daly, C. A.; Berquist, E. J.; Brinzer, T.; Garrett-Roe, S.; Lambrecht, D. S.; Corcelli, S.

A., J. Phys Chem. B 2016, 120, 12633-12642. 73.

Gierer, V. A.; Wirtz, K., Z. Naturforsch. A 1953, 8, 532.

74.

Perrin, F., J. Phys. Radium 1936, 7, 1-11.

75.

Moganty, S. S.; Baltus, R. E., Ind. Eng. Chem. Res. 2010, 49, 9370-9376.

76.

Moya, C.; Palomar, J.; Gonzalez-Miquel, M.; Bedia, J.; Rodriguez, F., Ind. Eng. Chem.

Res. 2014, 53, 13782-13789. 77.

Scovazzo, P.; Kieft, J.; Finan, D. A.; Koval, C.; DuBois, D.; Noble, R., J. Membr. Sci.

2004, 238, 57-63.

78.

Budhathoki, S.; Shah, J. K.; Maginn, E. J., Ind. Eng. Chem. Res. 2017, 56, 6775-6784.

36


Page 37 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 1. Parameters from the Triexponential Fits to the Anisotropy Decays Sample A1a t1 (ps) Bulk 0.11 ± 0.01 0.9 ± 0.1 SILM (pore) 0.13 ± 0.01 0.6 ± 0.1 SILM (polymer) 0.10 ± 0.01 0.6 ± 0.1 a Ai is the amplitude of each exponential.

A2a 0.08 ± 0.01 0.08 ± 0.01 0.03 ± 0.01

A3a 0.12 ± 0.01 0.14 ± 0.01 0.10 ± 0.03

t2 (ps) 6.5 ± 0.7 7.5 ± 0.9 6.1 ± 0.6

t3 (ps) 51 ± 1 90 ± 2 309 ± 119

offset 0 0 0.12 ± 0.03

Table 2. Parameters from Wobbling Analysis Sample

θin (deg)a

θc1 (deg)b

θc2 (deg)b

θtot (deg)c

τc1 (ps)d

τc2 (ps)d

τm (ps)d

Dc1 (10-2 ps-1)e

Dc2 (10-2 ps-1)e

Bulk

22.5 ± 1.5

30.3 ± 0.9

33.3 ± 0.8

48.7 ± 0.5

0.9 ± 0.1

7.4 ± 0.9

51 ± 1

8.2 ± 1.4

1.2 ± 0.1

SILM (pore)

16.6 ± 2.7

31.5 ± 1.3

30.4 ± 0.7

45.5 ± 0.4

0.6 ± 0.1

8.2 ± 1.1

90 ± 2

12.9 ± 2.3

0.9 ± 0.1

a

b

c

The inertial cone angle. θc1 and θc2 are the first and second diffusive cone half angles. The total cone half angle accounting for all three cones. dτc1, τc2, and τm are the decay times associated with the first and second diffusive cones and the final free diffusion, respectively. eDc1 and Dc2 are the first and second cone diffusion constants.

Table 3. SSD Parameters from RISD Fits Based on the Second-Order Stark Effect Model Sample IL Bulk IL in pores a

Fs, scalar correlation A1a τ1 (ps) 0.09 ± 0.02 20 ± 5 0.04 ± 0.01 17 ± 8

A2a 0.07 ± 0.02 0.17 ± 0.01

τ2 (ps) 77 ± 20 143 ± 11

Fv, vector correlation A3a τ3 (ps) 0.36 ± 0.01 57 ± 5 0.54 ± 0.01 113 ± 7

Ai is the amplitude of each exponential.

Table 4. Isotropic FFCF parameters Sample

Γ (cm-1)a

∆1 (cm-1)b

τ1 (ps)c

∆2 (cm-1)b

τ2 (ps)c

IL Bulk

3.2 ± 0.1

0.99 ± 0.02

7.2 ± 1.2

0.95 ± 0.05

52 ± 4

IL in pores

2.7 ± 0.2

0.98 ± 0.04

14.2 ± 1.9

1.12 ± 0.06

82 ± 12

a

Γ: homogeneous line width (FWHM). b∆i: inhomogeneous line width (standard deviation) of the ith component. cτi: decay time constant of the ith component.

37



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 44

Figure Captions Figure 1. Background-subtracted FT-IR of 13CO2 in (A) SILM, bulk EmimNTf2, (B) PES200,

and PES200 filled with mineral oil. The spectrum in red is obtained by subtracting the spectrum in bulk EmimNTf2 from the SILM spectrum. Figure 2. Population decay curves of the asymmetric stretch of 13CO2 in the IL (2277 cm-1 band)

and supporting membrane phases (2271 cm-1 band) of the SILM sample. The solid curves are the kinetic model fit described in ref 18. Figure 3. Anisotropy decay curves of 13CO2 in the bulk EmimNTf2 and the IL and supporting

membrane phases of SILM. The solid curves are triexponential fits to the data. Figure 4. 2D IR spectra of the asymmetric stretch of 13CO2 in SILM. (A): the spectrum at short

waiting time. The bands from both the IL and supporting membrane phases are elongated along the diagonal, the dashed line. (B): the spectrum at longer waiting time. The supporting membrane phase band at 2271 cm-1 is fairly elongated while the IL phase band at 2277 cm-1 looks almost round. The spectral diffusion is obtained from the change in shape of the spectrum with Tw. Figure 5. 2D IR CLS (normalized frequency-frequency correlation function) decay data of 13

CO2 (A) in the bulk EmimNTf2 and IL phase of SILM (2277 cm-1) from both the parallel and

perpendicular polarization configurations and (B) in the bulk EmimNTf2 and the IL (2277 cm-1) and supporting membrane phases (2271 cm-1) of SILM from the perpendicular polarization configuration. The solid curves are the second-order RISD fits to the data, except for the supporting membrane phase data which are fit with biexponential.

38


Page 39 of 44


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

absorbance (norm.)

A

absorbance (norm.)

B

CO2 in SLIM (1) bulk scaled (2) (1) – (2)

1.0 0.8 0.6 0.4 0.2

bend combination band

0.0

1.0 0.8

CO2 in pores empty pores min. oil (1) – (2) from A

0.6 0.4 0.2 0.0 2240 2250 2260 2270 2280 2290 2300 2310 frequency (cm-1)

Figure 1

39



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1.0

2277 cm-1 IL 2271 cm-1 polymer

0.8

P(t) (norm.)

Page 40 of 44

0.6 0.4 0.2 0.0 0

50

100

150

200

250

300

t (ps)

Figure 2

40


Page 41 of 44


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

0.30 SILM polymer SILM pore Bulk

r(t) - anisotropy

0.25 0.20 0.15 0.10 0.05 0.00 0

25

50

75

100 125 150 175 200

t (ps)

Figure 3

41



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 42 of 44

A 2285 T = 4 ps w m (cm-1)

2280 13CO in 2 the IL

2275

13CO 2

in the polymer

2270

B

2285 2265

Tw = 60 ps

m (cm-1)

2280

2275

2270

2265 2265

2270

2275

 (cm ) -1

2280

2285

Figure 4

42


Page 43 of 44


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

A

0.5 Bulk EmimNTf2 SILM (pore)

CLS(Tw)

0.4 0.3 0.2 0.1

B

0.0

0.7

CLS(Tw)

0.6 Bulk EmimNTf2

0.5 0.4

SILM (pore) SILM (polymer)

0.3 0.2 0.1 0.0

0

25

50

75

100

125

150

175

Tw (ps) Figure 5

43



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Polymer phase

Page 44 of 44

IL phase in pores

2240 2250 2260 2270 2280 2290 2300 2310 frequency (cm-1)

TOC figure

44


Carbon Dioxide in a Supported Ionic Liquid Membrane: Structural and

Recommend Documents